Abstract/Summary

Extended bilingualism can induce first language (L1) attrition, and lead late sequential bilinguals to deviate from monolingual peers in terms of L1 representation and processing (Schmid & Köpke, 2017). Various theoretical frameworks (e.g. the Interface Hypothesis, Sorace, 2011) have been proposed to explain and predict which L1 domains are vulnerable to attrition, but they still need more empirical evaluation. Moreover, these frameworks have rarely discussed an important issue, i.e. whether and how individual differences in cognitive abilities might explain individual variation in L1 attrition. Framed within the Interface Hypothesis (IH), this study investigated 14 L1 Mandarin-L2 English bilinguals, and examined whether L1 attrition was more likely to happen with the long-distance binding reflexive ziji, a structure at the “external” syntax-pragmatics interface, rather than perfective and durative aspect marking in simple declarative sentences, a structure at the “internal” syntax-lexicon interface. This study also examined whether individual differences in a cognitive ability, namely working memory (WM) capacity, might explain potential variations in L1 attrition. Both off-line and on-line measures were employed in a novel combination including eye-tracking, to assess the perception, production/ interpretation and processing of these structures. In line with the IH, the bilinguals only demonstrated L1 attrition in interpreting and processing ziji, which assumingly involve syntax and pragmatics. However, the IH cannot fully explain the specific patterns of L1 attrition observed in this study. Furthermore, the results partially supported our hypothesis that individual differences in cognitive abilities could explain variation in L1 attrition. The interaction between WM capacity and other factors explained variation in processing ziji, but not in interpreting ziji off-line. Nevertheless, this finding highlighted the possible role of cognitive factors in explaining variation in L1 attrition. We conclude with suggestions in how future research can be informed by taking a more a holistic, cross-disciplinary approach to L1 attrition.